Skip to Content

Register Login

Industry Online Support

Technical Forum

Navigation

How to switchover CPU when CP441 External fault lights up

Created by: Techy at: 3/26/2007 1:30 PM (10 Replies)

Rating (1)

Thanks 0

Result pages:: |; 1 | 2; |

11 Entries

Entries per page: 10 | all

3/26/2007 1:30 PM	Rate (0)
Techy Posts: 86 Rating: (2)	I have a S7 400H system with CP441 in each rack. I would like to switch the master CPU to standby CPU if there is cable breakage on CP441 ( I assume EXTF LED glows??) Any suggestions which SFC or system status I need to monitor & how to force switch the master CPU? Thanks, Techy
Techy Posts: 86 Rating: (2)
Suggestion To thank Quote Answer

3/28/2007 7:42 AM	Rate (1)
Aret Posts: 2348 Rating: (264)	1. You can read which CPU is master by using SFC 51. You can read diagnostic events from module using OB82,83,86 or SFC 51/59. 2. The only thing i can suggest to force CPU switching is following: define that H-system is in redundant mode. If yes and diagnostic tells, that CP is not working => change some bit in DB and call SFC46 (stop function). Then 2nd CPU starts. Unfrotunately can't suggest how to restart remote CPU. 3. Think the best way to make such system woulb be using Y-link and CP after Y-link.
Aret Posts: 2348 Rating: (264)
Suggestion To thank Quote Answer

3/28/2007 6:38 PM	Rate (0)
Techy Posts: 86 Rating: (2)	Thanks Aret, was able to get it working. Regards, Techy
Techy Posts: 86 Rating: (2)
Suggestion To thank Quote Answer

3/29/2007 7:15 AM	Rate (0)
Aret Posts: 2348 Rating: (264)	Welcome^) If you will find out how to make good switchover algorithm - let me know :^)
Aret Posts: 2348 Rating: (264)
Suggestion To thank Quote Answer

3/29/2007 11:58 AM	Rate (0)
Techy Posts: 86 Rating: (2)	Aret, CP441 calls OB82 whenver there is a fault. I put latching flags based on whichever CP441 was going into fault mode. Based on this flag I was switching the FB180 block calls to healthy CP441 in standby rack. Thanks, Techy
Techy Posts: 86 Rating: (2)
Suggestion To thank Quote Answer

3/30/2007 6:37 AM	Rate (0)
Aret Posts: 2348 Rating: (264)	Nice, thanks^)
Aret Posts: 2348 Rating: (264)
Suggestion To thank Quote Answer

4/5/2007 12:11 AM	Rate (0)
Late Platinum Expert Joined: 1/7/2006 Last visit: 11/17/2023 Posts: 2705 Rating: (318)	Hi techy, I don't like it to switch a CPU in STOP when the CP441 is not working. It is better to detect in OB82 that there is a problem and don't call the software any more for the faulty detected CP441. I couldn't see what you have done now but maybe you have already do this now.
	I hope that the answer will help, if you have still questions, don't hesitate to ask me. All the best, Late Never say never, think in solutions, not in trouble......
Suggestion To thank Quote Answer

4/5/2007 7:13 AM	Rate (0)
Aret Posts: 2348 Rating: (264)	Hi Late, i had similar project, where a lot of DP slaves were used (CP was another - but its not important). The probles is - where there is CP fault (not CPU) - then no switchover takes place and therefore system "thinks" all DP-slaves are gone. System's main goal was to obtain parameters from DP slaves 24/7 uninterruptably. That's why H-system was chosen. As there was no redundancy in CPs i couldn't make CPU work with other's CPU CP. So, as i said before, i think Y-link (or something like) is best solution^)
Aret Posts: 2348 Rating: (264)
Suggestion To thank Quote Answer

4/7/2007 11:58 AM	Rate (0)
Late Platinum Expert Joined: 1/7/2006 Last visit: 11/17/2023 Posts: 2705 Rating: (318)	Hi Techy and Aret, The CP443-5DX04 in Redundancy (with ET200M behind it as switched I/O) will work properly, when there is one error and other DP-Slaves are ok on the same Profibus, then it is no reason for the H-System to switch over. However that is also for the onboard Profibus DP no reason to switch over. Over the communication between the two CPUs the data from the other Profibus DP (onboard and as CP) is transfered to the other CPU who have no contact with the slave any more. With a CP443-1EX11 the same happens. For Serial cards there is no Redundancy managed by the system, or you must have it in the ET200M where the CP341 can be reached from both CPUs (if of course built up as switched I/O). But there is also NO help of the H-System when you want to place two CP341 that are doing the same. So you have to built normally the redundancy by software.
	Last edited by: Late at: 07.04.2007 11:59 I hope that the answer will help, if you have still questions, don't hesitate to ask me. All the best, Late Never say never, think in solutions, not in trouble......
Suggestion To thank Quote Answer

4/12/2007 2:53 PM	Rate (0)
Aret Posts: 2348 Rating: (264)	Late, everything you're talking is right. But we had some(actually many of them) Simeas P connected to profibus lines (CP443). If you insert Y-link you can get only 240 bytes of info through it => to 1 Y-link you can connect not more than 5 Simeas P. Price for system organized in this way was too high. So the solution was - take 3rd party device, acting like Y-link - I mean it switches if no watchdog counter increase came from master . This device had GSD file, but connecting only to 1 line (not Y-link like). So we had to configure 2 different networks, inserting each simeas twice (in each network). With this configuration it should be no redundant lines (as CPU "thinks"). So i have to take care myself for switching algorithm. If you'll same configuration is insane - i will agree with you :^)))
Aret Posts: 2348 Rating: (264)
Suggestion To thank Quote Answer

|

Share this page:

Share this page on...

Follow us on